Get the results of TWRI-funded research through technical and special reports. Find abstracts and full-text online for our reports.


Escherichia coli and Dissolved Oxygen Trends in the Upper Llano River Watershed, Texas (2001-2016)

M. Schramm, T. Broad, T. Arsuffi

Trends in Escherichia coli (E. coli) and dissolved oxygen (DO) at Upper Llano watershed water quality monitoring stations were evaluated for calendar years 2001-2016. E. coli concentrations and loads were evaluated with linear regressions and Weighted Regressions on Time, Discharge, and Season (WRTDS). DO concentrations were evaluated with linear regression and generalized additive mixed models (GAMMs). WRTDS and GAMMs were also used to remove effects of year to year variations of flow on E. coli and annual variations in mean temperature on dissolved oxygen. Linear regression indicated a statistically significant decrease in E. coli concentrations on the main stem of the Llano River immediately downstream of the the North and South Llano River confluence. Linear regression did not indicate statistically significant changes in DO concentrations at any sites in the the watershed. From 2001 through 2016 total E. coli loads decreased by 3.99 × 10⁴ million colonies per day on the North Llano and 1.34 × 10⁷ million colonies per day on the Llano River. Flow-normalized loads decreased by 1.27 × 10⁵ million colonies per day and 5.87 × 10⁶ million colonies per day on the North Llano and Llano rivers respectively. Estimated E. coli load reductions appear substantial; however, the estimations are considerably biased which decreases the certainty that estimated load reductions are indeed real. Despite uncertainty, it is reasonable to infer that land use practices and changes in the watershed have not contributed to increases in flow-normalized E. coli loads between 2001-2016. GAMMs estimate that mean DO concentrations decreased 8.8% on the North Llano and 8% on the Llano during the same time period. Despite the decrease, mean DO concentration remain well within state water quality standards for aquatic life established for the watershed. The GAMMs, as designed in this analysis, are only used only to substantiate underlying trends in DO concentration. Because covariates that are characteristic of eutrophication are not included in this analysis, the GAMMs do not provide causal inference into mechanisms behind the observed responses in DO. If future research is warranted, additional instream monitoring and experimentation would be required to quantify possible eutrophic responses to changes in nutrients and chlorophyll-a.

Back to Top